• The Journal of the Acoustical Society of Korea
• /
• v.21 no.5
• /
• pp.462-470
• /
• 2002

Wideband speech, characterized by a bandwidth of 50∼7000 ㎐, sounds more natural and intelligible, and is less tiring to listen to when compared to narrowband speech characterized by a bandwidth of 300∼3400 ㎐. Wideband speech coders, however, have not been as successful as the narrowband speech coders because of their higher bit rate. In this paper, we propose a new wideband speech coder which combines the European standard of a narrowband speech coder, i.e., GSM-EFR, and a transform coder using the discrete wavelet transform. The proposed wideband speech coder operates as follows input speech is first split into two subbands with equal bandwidth and the two subband signals are coded and decoded by each subband coder. A GSM-EFR is adopted as a lower subband coder and a subband coder with wavelet transformed speech is designed for a upper subband coder. The total bit rate of the proposed coder is 18.9kbps (12.2 kbps for lower band coder and 6.7 kbps for upper band coder), and informal listening test results have shown that the proposed coder has comparable speech quality to that of G.722 with 56 kbps.

• Phonetics and Speech Sciences
• /
• v.14 no.4
• /
• pp.67-75
• /
• 2022

This study compared voice range profiles (VRPs) of modal and falsetto register in 53 dysphonic and 53 non-dysphonic adult women with gliding vowel /a/'. The results shows that maximum fundamental frequency (F0_MAX), maximum intensity (I_MAX), F0 range (F0_RANGE), and intensity range (I_RANGE) are lower in the dysphonic group than in the non-dysphonic group. F0_MAX and F0_RANGE are significantly higher in falsetto register than modal register in both groups. I_MAX and I_RANGE are significantly higher in falsetto register in the non-dysphonic group, but those are not different between two registers in the dysphonic group. There was no statistically significant difference in minimum F0 (F0_MIN) and minimum intensity (I_MIN) between the two groups. Modal-falsetto register transition occurred at 378.86 Hz (F4#) in the dysphonic group and 557.79 Hz (C5#) in the non-dysphonic group, which was significantly lower in the dysphonic group. It can be seen that both modal and falsetto registers in dysphonic adult women are reduced compared to non-dysphoinc adult women, indicating that the vocal folds of dysphonic adult women are not easy to vibrate in high pitches. The results of this study would be the basic data for understanding the acoustic features of voice disorders.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.12C
• /
• pp.1262-1267
• /
• 2005

Elliptical basis function (EBF) neural network works stably under high-level background noise environment and makes the nonlinear processing possible. It can be adapted real time VAD with simple design. This paper introduces VAD implementation using EBF and the experimental results show that EBF VAD outperforms G729 Annex B and RBF neural networks. The best error rates achieved by the EBF networks were improved more than $70\%$ in speech and $50\%$ in silence while that achieved by G.729 Annex B and RBF networks respectively.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.369-372
• /
• 2018

This paper suggests non-parallel-voice-conversion network conversing voice between unmapped voice pair as source voice and target voice. Conventional voice conversion researches used learning methods that minimize spectrogram's distance error. Not only these researches have some problem that is lost spectrogram resolution by methods averaging pixels. But also have used parallel data that is hard to collect. This research uses PPGs that is input voice's phonetic data and a GAN learning method to generate more clear voices. To evaluate the suggested method, we conduct MOS test with GMM based Model. We found that the performance is improved compared to the conventional methods.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1993.06a
• /
• pp.242-246
• /
• 1993

예측신경회로망 모델은 다층 퍼셉트론을 연속되는 음성특징 벡터간의 비선형예측기로 사용하는 동적인 음성인식 모델이다. 이 모델은 음성의 동적인 특성을 인식에 이용하고 연속음성인식으로의 확장이 용이한 우수한 인식 모델이다. 그러나, 예측신경회로망 모델은 음운학적으로 유사한 음성구간에서의 변별력이 낮다는 문제점이 있다. 그것은 기존의 학습 알고리즘이 다른 어휘와의 거리는 고려하지 않고 대상어휘의 예측오차만 최소화시키기 때문이다. 따라서, 본 논문에서는 직접 인식오차를 최소화시키는 GPD알고리즘에 의해 유사어휘간의 거리를 고려하는 변별력있는 학습 알고리즘을 제안한다.

• Journal of Broadcast Engineering
• /
• v.9 no.1
• /
• pp.9-16
• /
• 2004

In this paper, we present a speech enhancement method as a pre-processor for voice commander under car environment. For the friendly and safe use of voice commander in a running car, non-stationary audio signals such as music and non-candidate speech should be reduced. Ow technique is a two microphone-based one. It consists of two parts Blind Source Separation (BSS) and Kalman filtering. Firstly, BSS is operated as a spatial filter to deal with non-stationary signals and then car noise is reduced by kalman filtering as a temporal filter. Algorithm Performance is tested for speech recognition. And the results show that our two microphone-based technique can be a good candidate to a voice commander.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.8
• /
• pp.901-906
• /
• 2015

We presented a method for Vad(Voice Activity Detection) using Bi-level HMM. Conventional methods need to do an additional post processing or set rule-based delayed frames. To cope with the problem, we applied to VAD a Bi-level HMM that has an inserted state layer into a typical HMM. And we used posterior ratio of voice states to detect voice period. Considering MFCCs(: Mel-Frequency Cepstral Coefficients) as observation vectors, we performed some experiments with voice data of different SNRs and achieved satisfactory results compared with well-known methods.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2015.07a
• /
• pp.418-420
• /
• 2015

본 논문은 베이지안 비음수 행렬 인수분해 (Bayesian nonnegative matrix factorization, BNMF) 기반의 음성 강화 기법에서 음성과 잡음 성분의 latent source 수에 따른 강화성능에 대해 서술한다. BNMF 기반의 음성 강화 기법은 입력 신호를 서브 신호들의 합으로 분해한 후, 잡음 성분을 제거하는 방식으로 그 성능이 기존의 NMF 기반의 방법들보다 우수한 것으로 알려져 있다. 그러나 많은 계산량과 latent source 의 수에 따라 성능의 차이가 있다는 단점이 있다. 이러한 단점을 개선하기 위해 본 논문에서는 BNMF 기반의 음성 강화 기법에서 최적의 latent source 개수를 찾기 위한 실험을 진행하였다. 실험은 잡음의 종류, 음성의 종류, 음성과 잡음의 latent source 의 개수, 그리고 SNR 을 바꿔가며 진행하였고, 성능 평가 방법으로 PESQ (perceptual evaluation of speech quality) 를 이용하였다. 실험 결과, 음성의 latent source 개수는 성능에 영향을 주지 않지만, 잡음의 latent source 개수는 많을수록 성능이 좋은 것으로 확인되었다.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.3
• /
• pp.137-143
• /
• 2006

The aim of this paper is to investigate the prosodic characteristics of Korean distant speech. Four speakers (2 males and 2 females) produced 36 2-syllable words in both distant-talking and normal environments. totaling 288 spoken 2-syllable words. The results showed that ratios of second syllable to first syllable in vowel duration and vowel energy were significantly larger in the distant-talking environment compared to the normal environment and f0 range also bigger in the distant-talking environment. In addition, 'HL%' contour boundary tone in the second syllable and/or 'L+H' contour tone in the first syllable were used in the distant-talking environment.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.08a
• /
• pp.339-343
• /
• 1998

Sinusoidal model 은 음성신호처리의 넓은 분야에 적용되고 있는 방법으로 고음질의 합성음을 생성해 낼 수 있고, 조작이 용이하다는 장점을 가지고 있다. 본 논문에서는 Analysis-by-synthesis/Overlap-add Sinusoidal model 이라는 방법을 이용하여 시간축 변환과 dam성 변환을 수행하였다. 특히 본 논문에서는 음질향상을 위하여 시간축 변환시에는 정적인 구간과 변화하는 구간을 구별하여 서로 다른 시간축 변환비를 이용하였고, 기존의 LPC 방법에 비해 스펙트럼 포락선을 보다 잘 추정하는 Improved Cepstrum을 이용하여 음정변환에 적용하였다. 또 서로 다른 문맥에서 얻어진 음성단위들을 결합할 때 생기는 위상차이를 극복하기 위하여, 기본주파수 성분이 일치하도록 시간축을 이동하여 합성하였다. 실험결과 본 논문에서 적용한 방법들을 통해 기존 방식에 비해 개선된 음질을 얻을 수 있었다.